LAUSR

To the Editor: We read with interest the article by Gavaud et al (1), published in a recent issue of Critical Care Medicine, on the potential interest of the respiratory variability of the pulmonary velocity-time integral (ΔPulmVTI) to predict fluid responsiveness. Although this is an interesting study, with a potential impact on patient care, we would like to discuss a few physiologic points which may help clinicians making their own mind about the usefulness of this variable. The correlation reported by the authors between ΔPulmVTI and the respiratory variability of the aortic velocity-time integral (ΔAoVTI) and its surrogate, pulse pressure variation (PPV), is at the least expected. Both ΔAoVTI and PPV reflect left ventricular preload dependency, due to a decrease in right ventricular (RV) ejection, that is, pulmonary VTI, during inspiration in spontaneously breathing patients or tidal ventilation in mechanically ventilated patients (2). Nevertheless, the impaired ability of the RV to eject blood into the pulmonary circulation during tidal inflation, making the left ventricle preload-dependent, can result from two opposite physiologic mechanisms. As in the study by Gavaud et al (1), some patients experience a decrease of the PulmVTI related to a decrease in systemic venous return due to increased intrathoracic pressure upon tidal inflation; this phenomenon is likely to be amplified by hypovolemia so that ΔPulmVTI represents an accurate surrogate for fluid responsiveness in this situation. However, what Gavaud et al (1) dismiss is that other patients may display the same PulmVTI decrease, but related to an increase in RV afterload rather that a decreased preload (3); this is especially true in case of RV failure, a situation where fluid expansion is useless and even deleterious (4). In other words, what Gavaud et al (1) observed in a very selected population of fluid responsive patients (90% of the 30 filled patients were responders, and the cardiovascular system was not tested in the remaining 20 patients) is not applicable in a less selected population including patients at risk for afterload effect, such as patients ventilated for acute respiratory distress syndrome or patients with RV failure, situations which were both excluded in the study by Gavaud et al (1). In practice, PPV, exactly as ΔPulmVTI, are warning signals that abnormal interaction between the RV and the ventilator occurs, but neither can differentiate RV preload from RV afterload effect (5). This differential diagnosis rather relies on a hemodynamic variable reflecting specifically the effect of increased intrathoracic pressure on the resistance to venous return, namely the respiratory variation of the superior vena cava. In conclusion, ΔPulmVTI is a useful variable obtained by critical care echocardiography to evaluate whether RV function is impaired or not by mechanical ventilation. However, it cannot be considered as a new and accurate predictor of fluid responsiveness, but rather as an updated 2.0 version of the “old” PPV, with the same clinical efficacy in selected (fluid responsive) patients, but the same limits in nonselected intensive care population. The authors have disclosed that they do not have any potential conflicts of interest.